On 06/29/2012 04:23 PM, Kamezawa Hiroyuki wrote: > (2012/06/28 19:54), Sha Zhengju wrote: >> This patch series provide the ability for each memory cgroup to have independent >> dirty/writeback page stats. This can provide some information for per-cgroup direct >> reclaim. Meanwhile, we add more detailed dump messages for memcg OOMs. >> >> Three features are included in this patch series: >> (0).prepare patches for page accounting >> 1. memcg dirty page accounting >> 2. memcg writeback page accounting >> 3. memcg OOMs dump info >> >> In (0) prepare patches, we have reworked vfs set page dirty routines to make "modify >> page info" and "dirty page accouting" stay in one function as much as possible for >> the sake of memcg bigger lock. >> >> These patches are cooked based on Andrew's akpm tree. >> > Thank you !, it seems good in general. I'll review in detail, later. > > Do you have any performance comparison between before/after the series ? > I mean, set_page_dirty() is the hot-path and we should be careful to add a new accounting. Not yet, I sent it out as soon as I worked out this solution to check whether it's okay. I can test the series after most of people agree with it. Thanks, Sha -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>